Spatiotemporal Data Modeling Based on XML

1 Introduction

In the past, spatial database and temporal database are two mutually independent research fields. From the 1990s, researchers gradually realize that there are relations between these two fields and begin to study the combined knowledge: spatiotemporal database (Pelekis et al., 2004; Vazirgiannis & Wolfson, 2001). As establishment of spatiotemporal data model is the core issue of spatiotemporal database, researchers have proposed many spatiotemporal data models (Akhter et al., 2012; Camossi et al., 2003; Jonathan et al., 2001; Pelekis et al., 2004). However, the topics of representing and querying spatiotemporal data have received little attention in XML, which has emerged as the lingua franca for data exchange, and even possibly for heterogeneous data representation. This leads us to the question of spatiotemporal data model based on XML.

Being one of the most important branches of database, spatiotemporal database has extensive applications (Baazizi et al., 2011; Chen & Revesz., 2003) and receives increasing attention (Cobb & Petry, 1998). As the purpose of spatiotemporal database is managing spatiotemporal data, researchers have made a great amount of effort in dealing with spatiotemporal data. To this end, how to accurately and effectively operate spatiotemporal data is the core issue of managing spatiotemporal data. Unfortunately, those proposed efforts are based on traditional database, which will generate new tuples and produce a great deal of redundancy (Chen & Revesz., 2003; Deng et al., 2009). Furthermore, the traditional database imposes strict restrictions on structure (Du et al., 2012) and data types (Chen & Revesz., 2003) of spatiotemporal data.

As the next generation language of the Internet, XML is rapidly emerging and playing an increasingly important role. In addition, XML has good expansibility and has great influence on data management (Franceschet et al., 2007) since it can remain tree structure of XML and can consider node or subtree as metadata. Furthermore, XML can augment each node with time labels to represent validation so that can coalesce meeting or overlapping time intervals into one time interval and then group them (Baazizi et al., 2011; Franceschet et al., 2007; Hadjieleftheriou et al., 2002). In that case, the advent of XML seems to provide a good way to solve the problem of operating spatiotemporal data.

Table 1 describes changing history of a spatiotemporal object using traditional database mode, which will generate a new tuple when any attribute value changes. Thus it will produce a great deal of redundancy. For example, the owner of o₁ is Jerry from 2000-01-01 to 2002-05-31. However, two tuples store it. What’s more, frequently coalescing tuples will occur when query deals with time and those coalesce is also complex in database.

Representation of spatiotemporal information using XML will solve the problems. We augment time labels to represent the validity of the element node or attribute node. Any meeting or overlapping time interval will coalesce into one time interval, and then they will be grouped as shown in Table 2. This nesting representation is hard to present in plane table, while XML just has the advantages of representing nesting objects. As shown below: